Current Project: machine learning for transport analysis and predictions
Joe is a lifelong NJ resident, growing up near Atlantic City and going to Princeton University for undergrad. His research centers around tokamaks, looking to bring together a variety of codes (both physics-based like TRANSP/ASTRA and a new machine-learning based model he developed for his first-year project) for predicting the way a plasma will evolve. We hope that combining a variety of predictions will give us a more robust and accurate estimate. One important end-goal of such research is a model that can be used in realtime control algorithms to guide plasmas to desired scenarios. Joe's second-year project was in developing such a realtime control algorithm, and he continues running related experiments at the US tokamak DIII-D.